Forest For Tomorrow Silviculture Strategies

Download Report

Transcript Forest For Tomorrow Silviculture Strategies

Model Application: input layers,
data management, and model run
sequence
Robin McKinley, Wildlife Infometrics Inc.
Objective



Illustrate the data layers necessary for
running the models;
Describe data management and obstacles;
and,
Provide an overview of the model run
sequence
Model Input Layers



23 original data sets are required for running the models
6 additional data sets are derivatives of model results,
prepared mid-model, and combined with the existing
resultant;
Original datasets come from the following sources:









DEM and DEM derivatives
TRIM features and derivatives
VRI
BTM and BTM derivatives
BEC
Digital Road Atlas
Land and Resource Data Warehouse
Wildlife Habitat Ratings
Salmon Occurrence Probabilities
Raw Data Processing
Classified into
model states
(DEM -> Elevation)
Spatial modeling
routine
(Solar Radiation)
Results classified
into model states
Netica Manager
Codes data inputs directly
from databases
Spatial processing
of model results
Classified into
model states and
reincorporated into
model inputs
Raw Input Data
DEM and DEM Derivatives




25 m DEM acquired from NCC
Data extracted from the DEM include: Aspect,
Elevation, and Slope;
Derived/modeled data include: Moisture Regime,
Topographic Roughness, Solar Radiation, and
Topographic Curvature;
DEM data was processed at 25 m, reclassified
into model input states, then resampled to 100
m
VRI


Polygon data acquired from the NCC
Processing steps for VRI:




Assign a unique id for each polygon
Merge areas together
Export table of unique ids and attributes
Convert to grid
BTM and BTM Derivatives


Baseline Thematic Mapping coverages acquired from
NCC
Processing steps:





Create a topology that identifies overlaps and gaps in the
multiple input layers;
Correct topological errors and merge layers together;
Assign a unique id to the polygons;
Convert to grid
Netica handles the raw input (BA: Bare Areas and IBS:
Ice and Bare Sites) while a spatial input layer is prepared
for PTHD: Proximity To Human-caused Disturbance
TRIM Derivatives


TRIM dataset of the study area provided
by NCC via DEA with Ministry of Forests
(Don Morgan)
Derivative data sets include: Proximity to
Major Rivers

Double line rivers were selected and buffered,
proximity was calculated and the results were
reclassified and converted to grid
Other Basic Input Layers




BEC
Wildlife Habitat Ratings: Grizzly Bear, Black
Bear, Wolverine, Lynx
Digital Road Atlas
Salmon Occurrence Probabilities
Modeled Input Layers



There are 2 mid-model spatial processing
steps which create additional data layers
Modeled input are derivative layers from
model results
Distance to dens (GUGU / LYCA); Distance
to cover (interception); Distance to forage
(MAPE); and Distance to predation risk
(RATA)
Data Acquisition

Data input gathering, management, and
preparation


Raw data was gathered from NCC, MoFR,
MoE, and ILMB
Data was sorted for modeling based on:
Simple stratification into node states (BEC);
 Required scripting modifications in Netica Manager
(VRI);
 Required modeling to derive node states (Solar
Radiation from DEM)

Netica Manager Build

Netica Manager (NM)



MS Access form that codes data inputs from
the spatially referenced input data into their
model node states
manages the case files for all BBNs where
classified values are exported to an ASCII file
for processing in Netica
imports the BBN model results and joins them
back to the spatial database
Spatial Layer Construction

Spatial layer construction, distribution, and
combination






Time intensive process;
Some data requires simple stratification (Aspect,
Elevation);
Some data requires database scripting in Netica
Manager (VRI);
Some data requires modeling for node state
stratification (Solar Radiation, Moisture Regime);
Input data clipped and distributed to 33 processing
units;
Input data combined into a resultant spatial grid for
processing and attributes exported to a database
Model Run Sequence
Netica Manager
.cas file
Run Netica
Models
RES A
Link to spatial
Spatial
processing
RES B
Model Outputs



For every species, probability of
occurrence was expressed as density (and
standard deviation) of animals
Output grids for the entire study area
were put together using a mosaic routine
Grids were combined to create a resultant
grid for a species